A New Approach of Combining Optical Mapping Algorithm with Adaptive Kalman Filter to Achieve Fast and Early Detection of Cardiac Arrests: A Parallel Implementation

This research aims to propose a new approach by combining the Optical Mapping Algorithm (OMA) with the Adaptive Kalman Filter (AKF) to improve the early detection of cardiac arrest. To improve the overall performance of the proposed approach and reduce the execution time significantly, a parallel implementation is suggested using the open-source computer vision (OpenCV) library tool and optimized for heterogeneous multi-core systems. The OpenCV library incorporates many image processing functions that are used to extract the distinctive features of the digital image to indicate the presence or absence of cardiac arrest in order to save the largest possible number of human lives. Experimental results confirm the efficiency of the proposed implementation; as for a loop with a sequence of operations repeated 64 times, the achieved speedup is 5X with an accuracy improvement of 96%. This is due to the exploitation of several optimization techniques, such as pipelining and loop unrolling. The accuracy of dark spot detection of blood vessels by combining the AKF with OMA was improved from 50% to 83.3% compared with the AKF approach alone. Thus, the proposed approach and its parallel implementation are very useful in early cardiac arrest detection to preserve human lives.